Add argument to skip filtering if so desired.

wukevin / babel

Deep learning model for single-cell inference of multi-omic profiles from a single input modality.

https://www.pnas.org/content/118/15/e2023070118

38 stars 16 forks source link

Add argument to skip filtering if so desired. #5

Closed rcannood closed 3 years ago

rcannood commented 3 years ago

This PR adds an argument which skips filtering. This allows training babel methods on small (synthetic) datasets. If filtering is not turned off, the dataset is completely empty after filtering.

wukevin commented 3 years ago

This is a good addition. One thing to point out: Under train_model.py, the --nofilter option is only handled if --shareseq and --snareseq are not specified. This seems reasonable, but it'd be better for the CLI help docs to be explicit about this. I will make a separate commit to address this after merging.

Thanks!